Seite - 539 - in Menschenrechte und Gerechtigkeit als bleibende Aufgaben - Beiträge aus Religion, Theologie, Ethik, Recht und Wirtschaft

© 2020, Vandenhoeck & Ruprecht GmbH & Co. KG, Göttingen ISBN Print: 9783847111658 – ISBN E-Lib: 9783737011655 replicated in AI are precisely the values that are already operative in society. Thus, the values that a society prioritises (whether these be efficiency, wealth accumulation,personal autonomy, equality, sharedbenefit), the explicit orun- acknowledgedbiases that it purveys (for example assumptions about the links betweenethnicityorraceandsocialresponsibility/mobility), thetrade-offs that asocietyaccepts(forexample,personalisationoverprivacy)andtheinequalities that it tolerates, will not only be embedded in AI systems, but will also be amplified through AI. Of course, AI brings new challenges and new ethical questions, however at itsmost fundamental level the deployment of reflexive/ self-learningalgorithms/systemsreplicateandmagnify thevaluesaroundwhich a society is already structured and amplify existing patterns of human deci- sion-making.TheonlydecisionsAImakesarethosethat it isdesignedtomake. 5. TrustworthyAIandValueSensitiveDesign InApril2018,theEuropeanCommissionpublishedaCommunication,whichset out a blueprint for a three-pronged approach to increase public and private investment inArtificial Intelligence, toprepareforsocio-economicchangesand toensureanappropriateethical and legal framework.Aspartof thisapproach, the Commission established the Artificial Intelligence High-Level Expert Working Group (AI HLEG), aiming to support the creation of the European Union’sstrategyonAI.Acorecomponentofthestrategyhasbeentosupportthe developmentofethical,human-centricAI.Themostrecentstageinthatstrategic processhasbeen the issuingof ‘EthicsGuidelines forTrustworthyAI’,16which setsout a framework for achievingethicalAI. Following in thepathof its 2015 ‘Opinion on (LegalQuestionsRelated to) Robotics andAI’17 the report is un- ambiguous in its conviction that the design anddevelopment ofAI should re- spect the fundamental values and the rights of freedom, autonomy, human dignity,non-discriminationandequality. In the2019Guidelines itdistils these fundamental values to four. These are the principles respect for human au- tonomy,preventionof harm, fairness andexplicability. Inadditionto enumer- atingtheethicalprinciplesandtheircorrelatedvaluesthatmustberespectedin thedevelopment,deploymentanduseofAIsystems, the ‘EthicsGuidelines’also provide guidance on howTrustworthyAI can be realised, by listing seven re- quirements thatAIsystemsshouldmeet inorder tobe trustworthy.Theseare: 16 EuropeanCommission,High-LevelExpertGrouponArtificialIntelligence,EthicsGuidelines forTrustworthyAI,Brussels2019. 17 EuropeanCommission, ExpertGrouponEthics, Opinionon (LegalQuestionsRelated to) RoboticsandAI,Brussels2015. DesigninganEthicalFuture:CanArtificial IntelligenceHelp? 539 Open-Access-Publikation im Sinne der CC-Lizenz BY 4.0